Automated behavioural response assessment to a feeding event in two heritage chicken breeds

•Video was recorded around a feeding event in two heritage chickens breeds.•Video was analyzed using a software-based method and an ethogram-based manual method.•The software method captured pre-feeding increases and post-feeding decreases in activity.•The ethogram data supported the findings of the automated analysis.•This study demonstrated an automated means applicable to chicken welfare monitoring.

Video was recorded for three consecutive days for two pens each of two breeds of heritage chickens, Rhode Island Red (RIR) and Barred Plymouth Rock (BPR) (34 birds/pen), before and following a feeding event. The feeding event occurred when a technician entered the pen to refill the food trough at 08:00 daily. Changes in overall activity were assessed using a software-based automated quantitative method, and changes in mobility, resting, comfort and feeding behaviour were scored manually. With the software method, activity was determined for 30 min pre- and post-feeding event, and was defined as the percent pixel change of the total arena per 30 s (a change in one pixel indicated a change in the use of an area of ∼0.69 cm2). Video recordings were also scan-sampled by a human observer every 30 s for 15 min pre- and post-feeding event, and behaviours were scored according to an ethogram. The ethogram data tended to support the findings of the automated analysis. Activity levels increased before and were at their highest following the feeding event. Breed differences were apparent, as BPR were more active but reduced activity faster than RIR. Resting, ground pecking and comfort behaviour were decreased post-feeding event for both breeds. Feather pecking was more prominent in RIR, but did not differ before or after the feeding event. In this study we have shown that an automated, software-based behavioural assessment method can be applied to chickens, and it is capable of detecting subtle changes in activity. In future, use of such an automated method has application in production systems in its ability to serve as an early warning system of health concerns by finding deviations from baseline activity.